Significant progress has been made in the rational design of immunogens for an HIV vaccine. Atomic-level structures of viral surface proteins in key functional states, or in complex with neutralizing antibodies, reveal important viral epitopes, providing targets for antibody elicitation. Structure-based computational methods can increase the efficiency, accuracy, and likelihood of success for immunogen design efforts. Specifically, through structural analysis, manipulation, and redesign, we expect a variety of computational techniques to be applied to generate immunogens that focus the immune response: (1) away from undesirably immuno-dominant regions; and (2) towards specific target epitopes. Such rationally designed immunogens are thus expected to enable the elicitation of broadly neutralizing antibodies capable of neutralizing a diverse range of HIV-1 isolates. We also planned to use computational algorithms to improve antibody properties, including enhancing neutralization potency and solubility, and reducing polyreactivity and immunogenicity, for therapeutic use